Intranasal Administration of a Polyethylenimine-Conjugated Scavenger Peptide Reduces Amyloid-β Accumulation in a Mouse Model of Alzheimer’s Disease

Peptide synthesis

All the peptides were prepared by the batch fluorenylmethoxycarbonyl (Fmoc)-polyamide method. The sequence of Aβ₄₀ is DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGVV. Different lengths of peptides were synthesized with V24 to ^DP mutations: V24P(10–40), V24P(1–28), V24P(13–36), V24P(16–33), and V24P(19–30). The N-terminal and C-terminal ends were not chemically modified. Fmoc-amino-acid derivatives (four equivalents) were purchased from Anaspec (Fremont, USA) and coupled on Wang resin that was preloaded with the C-terminal amino acid (one equivalent) using benzotriazole-1-yl-oxy-tris-pyrrolidino-phosphonium hexafluorophosphate (four equivalents) and 4.45% (v/v) N-methyl morpholine in dimethylformamide (DMF). The Fmoc cleavage step was performed using 20% piperidine in DMF. Side-chain deprotection and peptide cleavage from the resin were performed simultaneously by stirring the resin with a mixture of 9.4 mL of trifluoroacetic acid, 0.25 mL of water, 0.25 mL of ethanedithiol, and 0.1 mL of triisopropylsilane at room temperature for 1–2 h, at which point the resin was removed by passing the reaction mixture through a G2 glass funnel. The crude peptide was precipitated from the filtrate by the addition of three volumes of ice-cold methyl t-butyl ether (MTBE) and centrifugation at 2000 g for 15 min at 4°C, washed twice with MTBE, and dried under a vacuum. The precipitated peptide was purified by reverse-phase HPLC using a Vydac C18 column (10 mm×250 mm) and an acetonitrile–water mixture containing 0.1% trifluoroacetic acid. Peaks were analyzed on a matrix-assisted laser desorption ionization (MALDI) mass spectrometer, and those containing the desired product were lyophilized and stored at – 20°C. To synthesize the PEI-conjugated peptide V24P(10–40)-PEI, PEI was conjugated to the C-terminal carboxyl group of the peptide. To avoid interference of the N-terminal amino group on peptide–PEI conjugation, the N-terminal group of the peptide was acetylated in the final synthetic step using four equivalents of acetic anhydride instead of an amino acid derivative.

Circular dichroism (CD) spectroscopy

The peptide samples were dissolved in 20 mM of sodium phosphate buffer and 150 mM KCl (pH 7), then incubated at 25°C. After different incubation times, the samples were placed in a 1-mm cell, and the CD spectra between 195 and 250 nm were recorded on a J-715 CD spectrometer (JASCO, Japan). The band width was set to 2 nm, and the step resolution was 0.05 nm. Two scans were averaged for each sample.

Thioflavin T (ThT) binding assay

A stock solution of 2 mM ThT was prepared in 140 mM KCl and 100 mM sodium phosphate buffer (pH 7.5) and filtered through a 0.22μm Millipore filter. A fresh working solution was prepared by adjusting the dye concentration to 200μM. A 30-μL aliquot of sample was mixed with 30μL of 200μM ThT dye solution for 1 min at room temperature, then the fluorescence emission between 450 and 600 nm was measured in a 3-mm path-length rectangular cuvette on a FP-750 spectrofluorometer (JASCO, Japan) with excitation at 442 nm.

Transmission electron microscopy

The samples were deposited on carbon-coated 300-mesh copper grids (Nisshin EM, Tokyo, Japan), incubated for 3 min for absorption, and then washed with water. Negative staining was done with 2% uranyl acetate for 3 min. After air drying, the samples were viewed using a Hitachi H-7000 electron microscope (Hitachi, Tokyo, Japan).

Cell viability assay

Mouse N2a neuroblastoma cells (ATCC) were cultured in Dulbecco’s Modified Eagle Medium (DMEM; HyClone, USA) supplemented with 10% fetal bovine serum (FBS; HyClone, USA) in 5% CO₂ at 37°C. For the cell viability assay, the cells were harvested, suspended at a density of 350,000 cells/mL in DMEM, and 100μL was plated in each well of a 96-well CellBIND polystyrene microplate (Corning, USA). Because the cytotoxicity experiments lasted for up to four days, cell proliferation was blocked using medium with no FBS. The plates were then incubated at 37°C under 5% CO₂ for 24 h to allow the cells to attach. The peptides were dissolved in DMSO as 6 and 12 mM stock solutions. Five microliters of the stock solution (6 mM) was diluted with 95μL of PBS (20 mM sodium phosphate buffer, 150 mM KCl, pH 7.0), and the sample was immediately added to 900μL of fresh DMEM to give a peptide concentration of 30μM. To test for toxicity inhibition, equal volumes of Aβ₄₀ and peptide inhibitor stock solutions (12 mM) were pre-mixed and then diluted into PBS to make the final concentration of 30μM for each peptide. The diluted peptide solutions were pre-incubated for 24 h at room temperature with shaking (50 rpm) before being added to the cultures. The medium in the wells of the 96-well plate was replaced with 100μL of peptide-containing medium, and the plate was incubated for 48 h. Cell viability was then determined by using the MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) toxicity assay [27]. MTT (10μL of 5 mg/mL) in PBS was added to each well, and after incubation for 4 h, the medium was removed; the MTT crystals were dissolved in 100μL of 90% isopropanol, 0.5% SDS, and 40 mM HCl; and the absorption at 570 nm was measured. Cell viability was calculated by dividing the absorbance of wells containing peptide samples by that of wells without any added peptide, expressed as a percentage of the value of cells with no added peptide. The experiment was repeated four times, and eight replicate wells were used for each sample and control in each independent experiment.

Synthesis of the PEI-conjugated peptide

All peptides were synthesized by the batch Fmoc-polyamide method [28]. To generate PEI-conjugated V24P(10–40), PEI was conjugated to the C-terminal carboxyl group of the peptide, and the N-terminal group of V24P(10–40) was acetylated to avoid dimerization or cyclization of the peptide during the PEI-conjugation reaction and to provide protection against exopeptidases. Acetylated V24P(10–40) (4.8 mg) dissolved in 4 mL of dimethyl sulfoxide was slowly mixed with 240μL of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (600 mM in 0.1 M MES, 0.5 M NaCl, pH 6), and then 240μL of N-hydroxysuccinimide (1200 mM in 0.1 M MES, 0.5 M NaCl, pH 6) was added. The mixture reacted at room temperature for 30 min with gentle shaking (70 rpm). PEI (288μL) was added, and the mixture was incubated overnight at room temperature with gentle shaking (70 rpm). The PEI-conjugated peptide, V24P(10–40)-PEI, was then separated from the unreacted PEI and V24P(10–40) by reverse-phase HPLC.

Animal experiment

The animal experiments were approved by the Institutional Animal Care and Use Committee of the Academia Sinica. APP/PS1 transgenic mice (B6C3-Tg(APPswe,PSEN1dE9)85Dbo/Mmjax), purchased from Jackson Laboratories (USA), were bred and genotyped as described [25, 29]. The mice had access to food and water ad libitum and were kept on a 12 : 12 h light–dark cycle. PEI and V24P(10–40)-PEI were dissolved at a concentration of 400μM in 100 mM NaH₂PO₄, 138 mM KCl (pH 5), and 4-month-old mice were given 2.5μL of PEI (as control) or V24P(10–40)-PEI in each nostril six times per week for the indicated period.

ELISA assays for Aβ₄₀ and Aβ₄₂

Concentrations of Aβ₄₀ and Aβ₄₂ in mouse brain homogenate were measured using ELISA kits (Invitrogen, MD, USA) according to the manufacturer’s directions. Briefly, the cortical or hippocampal tissue from PEI-treated or V24P(10–40)-PEI-treated APP/PS1 mice was weighed and homogenized at 4°C in the cell extraction buffer provided in the kit and supplemented with a protease inhibitor cocktail (Sigma, St. Louis, USA). The homogenates in Eppendorf tubes were then centrifuged at 13,000 rpm at 4°C for 10 min, and the concentration of protein in the supernatant was measured using the microBCA protein assay (Thermo, IL, USA). To perform the ELISA, the supernatants were diluted 10-fold, and the Aβ₄₀ and Aβ₄₂ concentrations were normalized to the protein concentration and expressed as ng/mg of protein.

Thioflavin S staining

Paraformaldehyde-fixed brain sections were applied with 1% (w/v) Thioflavin S (ThS) solution for 10 min at room temperature protected from light, and then washed with 80% ethanol and water to remove excessive dye. ThS-positive signals were visualized under an epi-fluorescence microscope; the plaque number, plaque area, and plaque size were analyzed by ImageJ (NIH, Bethesda, MD, break USA).

[¹¹C]PiB synthesis

[¹¹C] methyl bromide was produced by the multi-pass bromination of [¹¹C] methane. Subsequently, [¹¹C] methyl bromide was eluted from a trap and converted to [¹¹C] methyltriflate by passing a preheated silver triflate column. [¹¹C] methyltriflate was carried by a helium stream (20 mL/min) into 350μL of anhydrous methylethylketone containing 1.5 mg of 2-(4’-aminophenyl)-6-hydroxybenzothiazole. After the trapping was finished, the reaction mixture was heated at 75°C for 2 min, and then 0.4 mL of HPLC mobile phase was added to the reaction mixture for HPLC purification. HPLC purification was performed on a Waters Bondapak column (10μm, 7.8 mm ID × 300 mm) using a mobile phase of acetonitrile/0.01M H₃PO₄ (40/60) at a flow rate of 5.0 mL/min. The radioactive fraction corresponding to [¹¹C]PiB was collected in a bottle containing 30 mL of pure water and passed through a C18 Sep-Pak Plus cartridge, then washed with 10 mL of pure water, eluted with 1 mL of ethanol and 10 mL of sterile normal saline, and passed through a 0.22-μm sterile filter for quality analysis and the animal experiments. Radiochemical purity was greater than 99%, as determined by analytical HPLC. The specific activity was 152±52 GBq/μmol at the end of synthesis.

In vivo microPET

PiB was generated using [¹¹C] methyltriflate by a previously described method with minimal modification [30]. PET scans were performed using a Triumph preclinical tri-modality (LabPET/X-SPECT/X-O CT) imaging system (TriFoil Imaging, USA). The mice were kept warm with a heating lamp before scanning. After induction with 2.0% isoflurane, the mice were placed with their heads in the center of the field of view, fixed in the prone position, then freshly synthesized [¹¹C]PiB (36.7±2.6 MBq; volume <0.25 mL) was injected via the tail vein. After 20 min, static data acquisition was performed for 20 min in 3D list mode with an energy window of 350–650 keV. The emission data were normalized and corrected for the tracer decay time. All list mode data were sorted into 3D sinograms, which were then single-slice Fourier rebinned into 2D sinograms. Summation images from 20–40 min after [¹¹C]PiB injection were reconstructed using a MLEM algorithm.

All imaging data were processed and analyzed using the PMOD 3.5 software package (PMOD Technologies, Zürich, Switzerland). The PET image dataset was converted to an absolute measure of radioactivity concentration (kBq/cc) using a phantom-derived calibration factor before being normalized to the injected dose of [¹¹C]PiB and the body mass of the animal. Static PET images were co-registered with the mouse T2-weighted MRI brain atlas based on PMOD as anatomic reference. Image origins were set to bregma (0, 0) according to the MRI atlas, which was also used for VOI definition. [¹¹C]PiB uptake in the mice’s cortex, hippocampus, amygdala, and olfactory bulb was evaluated. Standardized uptake values were obtained for each VOI by dividing the mean [¹¹C]PiB activity by injection dose and body weight (g). Thereafter, the regional [¹¹C]PiB uptake in the target region was normalized to [¹¹C]PiB uptake in the cerebellum, which was used as the reference region [30, 31].

Peptide design

Previously people designed anti-amyloid peptides with the ability to bind Aβ monomer, oligomer, or amyloid fibrils in order to prevent Aβ amyloid deposition in the brain. However, it has been reported that several peptides which can inhibit Aβ fibrillization and reduce Aβ toxicity actually formed big aggregates [11, 13]. Moreover, van Groen et al. reported that a peptide, named D1, has Aβ binding affinity but can neither reduce Aβ amyloid deposition in the mouse brain nor retard cognitive declines of the mice [32–34]. These data suggest that anti-amyloid is not necessarily anti-aggregation. Our previous results showed the V24P peptide can form non-toxic aggregate with Aβ₄₀ [23]. The aim of this work is to find out the minimal sequence required to achieve the same effect, i.e., 1) form amorphous aggregates with Aβ and 2) prevent Aβ from forming toxic aggregates. Several mutated Aβ₄₀ peptides with different N- and C-terminal truncations were designed and synthesized (Table 1). We first examined their structural properties by CD spectroscopy, a Thioflavin T (ThT) binding assay, and transmission electron microscopy (TEM), and then we examined their effects on reducing Aβ₄₀ toxicity using a cell viability assay in mouse neuroblastoma Neuro2a (N2a) cells.

C-terminal-truncated peptide V24P(1–28)

According to the structural model of the Aβ₄₀ fibrils [35], K28 is located at the beginning of the second β-strand. Without the C-terminal hydrophobic tail following K28, V24P(1–28) has greater hydrophilicity compared with Aβ₄₀. V24P(1–28) was dissolved in the incubation buffer (20 mM sodium phosphate buffer, 150 mM potassium chloride (KCl), pH 7) and incubated at 25°C. At different time points (day 0 was the time immediately after sample preparation), its CD spectra and fluorescence spectra after ThT binding were recorded. The CD spectrum of the 30μM V24P(1–28) solution was typical of a random coil structure and was identical at all tested time points (0–12 days; Fig. 2a, left); consistently, the measurement of fluorescence showed no ThT binding (Fig. 2b, left). When the peptide concentration increased to 60μM, V24P(1–28) remained as a random structure from days 0–3, but gradually formed a β-sheet structure (Fig. 2a, right), as shown by the increased fluorescence (Fig. 2b, right). The TEM image showed that 60μM V24P(1–28) formed amyloid fibrils (Fig. 2e, left). The fibrils were straight and laterally associated, in contrast with the twisted morphology commonly reported for Aβ₄₀ fibrils [23, 37].

N-terminal-truncated peptide V24P(10–40)

The N-terminal region of Aβ₄₀ is hydrophilic and was reported not to be involved in the amyloid cross-β structure [35]. To verify whether this hydrophilic segment was important in the design of peptide inhibitors or not, an N-terminus truncated peptide without the first nine residues, V24P(10–40), was chemically synthesized and dissolved in the incubation buffer. In this study, 30μM V24P(10–40) exhibit a characteristic β-sheet signal, evidenced by negative ellipticity at 218 nm in its CD spectra (Fig. 2c, left) and consistent fluorescence emission at 487 nm in the ThT binding assay (Fig. 2d, left). At a concentration of 60μM, more β-aggregates were formed (Fig. 2c and 2d, right). Compared with the finding in our previous study, 30μM V24P adopted a random coil structure and the formation of amyloid-like β-aggregate was promoted by increasing peptide concentration [23]. These data showed that V24P(10–40) was more prone to aggregation than V24P. The TEM image proved that V24P(10–40) formed amorphous aggregates (Fig. 2e, right).

Structural studies on shorter peptide

In order to determine the shortest sequence for self-aggregation, we designed another three peptides with shorter lengths, V24P(13–36), V24P(16–33), and V24P(19–30), and examined their CD spectra and fluorescence spectra after binding ThT immediately after dissolved at concentrations of 30, 60, and 90μM.

The CD spectra showed that V24P(13–36) formed a random coil structure at 30μM (Fig. 3a, top) and the ThT fluorescence spectra showed that β-aggregates formed at 90μM (Fig. 3b, top). V24P(13–36) lacked several hydrophobic residues (Y10,V12,V39,V40) present in the V24P(10–40), and thus required a higher peptide concentration (90μM) to form β-aggregates compared with V24P (60μM) or V24P(10–40) (30μM).

In contrast, the CD spectrum of V24P(16–33) showed a strong peak at 204 nm and a trough at 228 nm (Fig. 3a, center panel), a pattern indicative of a type I β-turn [38, 39], while the weak fluorescence of 90μM V24P(16–33) suggested that only a small amount of β-aggregates formed (Fig. 3b, center).

V24P(19–30) showed a random coil structure. When the peptide concentration was 60μM or higher, the CD spectra showed a peak at around 220 nm (Fig. 3a, bottom), similar to that of an extended 3₁₀ helix or a poly(Pro) II helix [39], while no fluorescence emission at 487 nm was observed at any concentration (Fig. 3b, bottom). The data suggested that residues L17, V18, I31, and I32 are important for peptide aggregation and responsible for ThT binding in the V24P(16–33) aggregates.

Selection of a scavenger peptide

Our selection strategy is based on the aggregation propensity of the peptide inhibitor and its ability to reduce Aβ toxicity. We looked for a peptide which can interact and form non-toxic aggregates with Aβ. Length-shortening strongly increased peptide solubility. V24P(13–36) and V24P(16–33) could still form β-aggregate but only at a very high peptide concentration (90μM). The data shown in Figs. 2 and 3 suggest that only V24P(10–40) has the greater aggregation propensity than V24P to form non-fibril β-aggregate. Figure 4a shows that all the ^DP-containing peptides tested were much less cytotoxic than Aβ₄₀, suggesting that V24 is an important residue for Aβ₄₀ to form toxic species. When co-incubating V24P(10–40) and Aβ₄₀, V24P(10–40) could co-aggregate with Aβ₄₀ (see Supplementary Figure 1). Therefore, we surmised that V24P(10–40) attracts Aβ₄₀ and, together, they form non-toxic aggregates. Indeed, in the cytotoxicity assay, co-incubating Aβ₄₀ with V24P, V24P(10–40), and V24P(13–36) in a 1 : 1 molar ratio resulted in significantly higher cell viability (Fig. 4b). Among all these ^DP-containing peptides, V24P(10–40) was the most effective at reducing Aβ₄₀ toxicity. Moreover, the interaction is target-specific because V24P(10–40) could not inhibit fibril formation of prion peptides (see Supplementary Figure 2). Therefore, V24P(10–40) was chosen as the scavenger peptide in the subsequent animal study. V24P(10–40) contains the aggregation-prone LVFFA sequence and the GxxxG motif, both of which might contribute to its great ability of diverting Aβ association pathway [16].

Scavenger peptide V24P(10–40) decreased Aβ accumulation in the brains of APP/PS1 transgenic mice

In the design of peptide inhibitors, D-form amino acids, end capping, and methylation of amide hydrogens are often used to combat digestion by exopeptidases and endopeptidases in the serum thus extending the lifetime of the peptide in vivo [9, 15]. Designing peptide inhibitors that can cross the blood-brain barrier (BBB) is another challenge. Modification with putrescine, a naturally occurring polyamine, has been used to increase the ability of the designed peptide D-YiAβ11 to cross the BBB [20]. Several reports have also shown that polyethylenimine (PEI) cationized proteins have increased ability to cross the cell membrane [40–45]. Loftus et al. [46] demonstrated that PEI-conjugated green fluorescent protein can enter the brain when given intranasally, as the nerve cells of the olfactory epithelium project into the olfactory bulb of the brain, providing an excellent way of bypassing the need to penetrate through the BBB. We therefore added PEI to the C-terminus of our scavenger peptide V24P(10–40) and named the peptide V24P(10–40)-PEI. Although the modification of PEI attenuated the aggregation propensity of the peptide (Supplementary Figure 3), V24P(10–40)-PEI still had the ability to interact with Aβ and reduce Aβ toxicity (Supplementary Figures 3–5). The aggregates formed of co-incubated V24P(10–40)-PEI and Aβ₄₀ was more vulnerable for neprilysin degradation, as we proposed (Supplementary Figure 6).

We administered V24P(10–40)-PEI or PEI alone as nasal drops, to both nostrils of 4-month-old APP/PS1 transgenic mice (1 nmole or 4μg per nostril) six times per week, then sacrificed the mice four months later and analyzed the Aβ content of their brains by ELISA. As shown in Fig. 5a and b, the mice treated with V24P(10–40)-PEI clearly had less GdnHCl-extracted Aβ₄₀ and Aβ₄₂ levels in their cortex and hippocampus. The level of Aβ₄₀ and Aβ₄₂ of peptide-treated APP/PS1 mice was reduced to 28% and 60% of the control, respectively, in the hippocampus (Fig. 5a) and by 13% and 68% in the cortex. The amount of Aβ₄₂ was lower than that of Aβ₄₀. We surmised that some plaques might not be fully dissolved in GdnHCl extraction. Using ThS to stain the amyloid plaques in the brain tissue, we found that plaque number and plaque area were also reduced (Fig. 5c-e). The plaque number was decreased to 64% and 47% of the control group in the hippocampus and cortex, respectively. The percentage of area covered by plaques was decreased to 50% and 35% of the control group in the hippocampus and cortex, respectively. As for the effect on cognitive ability, reduction on escape latency in Morris water maze was observed on day 2 only, so the effect could not be clearly judged in this experiment (Supplementary Figure 7). Because we only tried one dosage in this animal experiment, we surmised that higher dosage might be required to stop the cognitive decline.

The reduction of amyloid plaque accumulation was quantified using micro positron emission tomography (microPET) with the radiotracer ¹¹C-labeled Pittsburgh compound B ([¹¹C]PiB), which binds to Aβ plaques. The APP/PS1 mice were treated with V24P(10–40)-PEI or PEI (2 nmole per day, six days a week) by intranasal administration from the age of 4 months to 12 months. The microPET images (Fig. 6a) taken at the end of treatment showed that the peptide treatment significantly decreased Aβ plaque deposition. Quantification of [¹¹C]PiB levels revealed that peptide treatment resulted in significant reductions of amyloid deposition to 89%, 81%, 73%, and 72% of the control group, in the cortex, hippocampus, amygdala, and olfactory bulb, respectively (Fig. 6b). The reduction of Aβ plaque deposition obtained by microPET is less than that observed by ThS staining. It might be because microPET was done in old mice which have more severe Aβ accumulation.